H₂S is a colorless, highly flammable and toxic gas, which in low concentration has an offensive odor similar to that of rotten eggs. As the Threshold Limit Value (TLV) for H₂S is 10 ppm, there is a need for its detection in either ppm or sub ppm level [1-3]. α-Fe₂O₃, an intrinsically n-type semiconductor has been widely exploited as a gas sensitive material owing to its good structural and thermodynamical stability along with resistance to photocorrosion. In the present work, H₂S sensing characteristics of Fe₂O₃ in both thin film and conventional forms have been investigated and compared. Our results indicate that thin film based sensors exhibited better response towards H₂S in comparison to that of pellets.

Fe₂O₃ thin films were prepared in two steps. In the first step, Fe films with ~100 nm thickness were deposited onto pre-cleaned (0001) Al₂O₃ substrates using electron-beam evaporation under a base vacuum of ~1.3 × 10⁻⁴ Pa. A high purity (99.99%) Fe powder was cold pressed in the form of a pellet and used as source material. These films were subsequently subjected to thermal oxidation at 800˚C in an oxygen environment (O₂ flow: 50 sccm) for 2 h. Bulk sensors in the form of pellets were prepared by cold pressing commercial Fe₂O₃ powder followed by sintering at 800˚C for 2 h. The response characteristics were recorded as a function of temperature, gas (H₂S, C₂H₅OH, NH₃, CH₄, CO, CO₂, NO, and Cl₂) and their concentrations using a static setup as described elsewhere [4]. For this purpose, the sensor was mounted on a Pt-100 based heater assembly in a leaktight glass chamber (net volume: 500 ml). Electrical contacts were made by thermally depositing two Au pads (120 nm thick). A desired concentration of the test gas in the chamber was achieved by injecting a known quantity of the test gas using a micro-syringe. The sensor response was calculated using the relations:

where G_g and G_a are conductance in the presence of test gas and air, respectively.

Figure 1 shows the X-ray diffraction (XRD) pattern ob-

tained for Fe₂O₃ thin film and bulk samples. All the peaks could be assigned to the single phase of Fe₂O₃ (hematite). The least-square fitting of the pattern indicated hexagonal rhombo-centred cubic unit cell structure with lattice parameters a = b = 5.028 Å, c =13.73 and α = β = 90˚, γ = 120˚, which are in agreement with the reported values [JCPDS card no. #79-0007].

Figure 2 is a plot of % response recorded for both the sensor films as a function of operating temperature towards 10 ppm of H₂S. It is clearly evident that thin films and bulk sensors exhibit response maxima at 250˚C and 200˚C, respectively. At all the temperatures the response of thin films is higher than that of bulk samples. For instance, thin film sensors show, response maxima of 262% in comparison to 112%, exhibited by bulk samples towards 10 ppm H₂S.

A Typical response curves for both the sensor films towards 10 ppm H₂S is shown in Figure 3. The conductance of the sensors increases on exposure to H₂S indicating an n-type response. Thin film sensors exhibited a maximum response towards H₂S. A baseline drift was observed for bulk sensors.

Figure 4 shows the selectivity histogram recorded upon exposure to 10 ppm of various reducing and oxidizing test gases. It is clearly evident that both the sensors are highly selective towards H₂S. A negligible response towards all the other interfering gases was observed.